Dendrite build up, killing batteries

spinningmagnets

Moderator
Staff member
Joined
Dec 21, 2007
Messages
12,951
Location
Ft Riley, NE Kansas
I've known for some time that when dendrites build up inside batteries, they can lead to a slow drain "killing" a cell in a group by going to zero volts) or if its bad enough, a cascading short leading to a battery fire.

I'm sure there's more to it, but...I came across an interesting statement on the web, and I have emailed the writer to get a link to the research (if possible). Don't make important decision based on an anonymous poster on the web, but...

"...an interesting discovery on dendrites turned up unexpectedly with Nissans experiment on Leaf owners, they gave half the study group a fast charger and half a normal one for a year, then took the batteries away for investigation. The slow-charged batteries had more growth. It seems that high-charge currents grow them one way, high-discharge the other, so...if you charge the same as you discharge, you would have near zero net growth..."
 
I know this was particularly true for old Ni-Cd batteries. Slow charging would build dendrites resulting in high self-discharge. A good blast of high current charge would blow away the dendrites like a fuse and the cell would be better for a while. The dendrites can puncture the separator and leave a hole that doesn't go away, so these cells would tend to go bad again over time.

There's probably more to it, but the basic idea may have some validity.
 
I recall when I was reading about working on old cars with points. There is a capacitor in the distributor circuit, and if it dies the manuals said its vital to get the exact amount of capacitance when buying the new capacitor.

If it had too much capacitance, ionized material from one of the physical contacts would erode away, and then build up on the other.

If there was not enough capacitance, the erosion/buildup was reversed.

Since it's often hard to get cheap parts that are precisely as advertised, it was common to file the faces of the contacts flat, and then se-set the gap during the "tune up".

I'd have two sets of points, and during the tune up, I'd simply swap them to get it over faster, plus to have a spare set. Once the car was running, could file the contact faces at my leisure.

For some reason, the flow of "dendrite" buildup flowing one way or the other stuck with me. I must confess I don't understand it, but this caught my eye.

Thanks, Richard, for sharing the info on NiCd's. We may be onto something here.
 
I remember using a large capacitor to blast a shorted cell back to life. Sometimes it worked sometimes it didn’t.

I haven’t heard of this being done with lithium batteries.
 
I don't fully get it, but if its true...
fechter said:
Slow charging would build dendrites resulting in high self-discharge. A good blast of high current charge would blow away the dendrites like a fuse and the cell would be better for a while.
Presumably, that means if you have a regen setup, each regen is a relative blast of high current charge, relative to whatever slower charging setup you may have (if it's correctly designed around the battery). So, regen good?

It feels like such a crapshoot anyway. I read all these threads on battery longevity, they feel like diet fads from the 90's: Regen good for batteries! Slow charge bad! Check out this new charging protocol that's sure to trim up your ebike's charge level is 6 weeks flat, or your money back!

So much info that can pro and con virtually any previously held notion you may have on batteries.
 
At a guess (with no personal experience / research of lithium dendrites), it probably also depends on the specific cell design--different chemistries, electrolytes, and even physical designs (cylindrical, prismatic, pouch) probably make a difference in response to various usages.


I do recall something about dendrites in a cell (of a chemistry I can't recall) puncturing cell layer separators, so that even if the dendrite was removed via whatever method, there was still a problem because of the separator damage; IIRC the dendrite could grow back there...but I cant' find the link to that thread now.
 
Dendrites are similar to disease that grows dependent on battery use over time. But you can slower that internal process to get more life time & more healthy life time.

There are some technics to slow dendrites grow

- dont charge battery more than 90% /discharge less than 20% (means for a 18650 lithium cell minimum aprox 3,3V and top 4,1V )
this will also double life cycles of cells just charging up 90% , and if you charge only up 80% arround 4V will tripe or x4 more cycles life time

- charge battery in small packets of charge (a full time charge will cause more stress/degradation to battery) example , if you go for a ride in your ebike for only 10km, just charge some few percentage. U dont need to be fully charged all time!

- using battery around center point voltage. means 3,6V or 3,7V nominal per cell. this requires some mental thinking..
Imagine a ebike 100km range, needs 20km's ride every day and every day charging. Soo you only need to charge a small packet of 20% every day around battery center point. for a 42V ebike battery the central point is 36V-37V , this means aprox 50% charged , some conclusion tells that U only need SOC state of charge interval from 40% to 60% , so just charge your battery from 40%- 60% every day!
This way battery cycles may increase x10. Battery is limited by chemicals internal natural life time. maybe lasts 10-15years.
Good quality batteries last 5years or more. I've seen some Sony 18650 cell from laptop with 12years performing good!

- Use the lower IR zone window. this means from 3,3V to 4,0V interval voltage. check discharge curve graph (high inclination zone means more IR) lower than 4,0V curve tend to stablize until arround 3,3V and drops.
This means U should use only aprox 70% of your battery capacity and leave the other 30% at rest, only charge battery every time up to 70%-80% never more. This will reduce few range but life time of cells will increase a lot!

-Temperature impacts internal chemistry reactions and harden dendrite crystals, so at low temperature use , do a pre battery heat, do not accelerate EV a lot at the beginning to prevent high currents flow. Ideal lithium cells functional temperature is around 30-40 degrees celcius (Fahrenheit ?? ) Similar behavior is a gas car, you should not play full throttle when motor is cold.. this cause degradation/damage to car motor, motor needs to pre heat similar to batteries to output full performance and less degradation!

PS: to avoid thinking about battery levels and percentages, volts, there is a quick way to limit charger top charging point.
Using diodes voltage drop effect. for example ebike charger 42V, add diodes with 0,7V drop at positive to have output voltage charge lower value. one diode equals 0,7V, soo for a 80% battery charger you need 3diodes in series (check current supported by diode to be safe at charge current) , with 3 diodes series you should have a 80% / 40V charger. do some test for results or add more series diodes.

You all, have a nice day!
 
Do i see someone spreading false rumors?

"dont charge battery more than 90% /discharge less than 20% (means for a 18650 lithium cell minimum aprox 3,3V and top 4,1V )
this will also double life cycles of cells just charging up 90% , and if you charge only up 80% arround 4V will tripe or x4 more cycles life time" is not proven with Science from what i know.

I made several videos debunking this:
Will your Li-ion batteries last twice as long if you charge them to 4.1V? https://www.youtube.com/watch?v=l0a6kikmS98
I Get Medieval on the science behind 4.1V DOD Research: https://www.youtube.com/watch?v=HhivZn_u5c0
Does Limiting DOD lead to longer life? - Lithium battery myths debunked Part III: https://www.youtube.com/watch?v=eGSJ64Cl_iY

If someone has any research/proof iäll gladlly look at it, i could be true, but please do not spread this rumour without proof.

Dendrites/SEI forms with fast charge, use and age.
 
As a leaf owner this is good information. Charge at level 2 and worried that I should be slow charging to prolong battery life.
As far as ebike batteries we should be charging at upper range of amps that the battery can handle?
 
The Battery Doctor said:
Do i see someone spreading false rumors?

"dont charge battery more than 90% /discharge less than 20% (means for a 18650 lithium cell minimum aprox 3,3V and top 4,1V )
this will also double life cycles of cells just charging up 90% , and if you charge only up 80% arround 4V will tripe or x4 more cycles life time" is not proven with Science from what i know.

I made several videos debunking this:
Will your Li-ion batteries last twice as long if you charge them to 4.1V? https://www.youtube.com/watch?v=l0a6kikmS98
I Get Medieval on the science behind 4.1V DOD Research: https://www.youtube.com/watch?v=HhivZn_u5c0
Does Limiting DOD lead to longer life? - Lithium battery myths debunked Part III: https://www.youtube.com/watch?v=eGSJ64Cl_iY

If someone has any research/proof iäll gladlly look at it, i could be true, but please do not spread this rumour without proof.

Dendrites/SEI forms with fast charge, use and age.
Please provide your curriculum vitae for you PhD in related science?
 
The Battery Doctor said:
Dendrites/SEI forms with fast charge, use and age.

This is sort of the opposite of what the guys testing the Leaf batteries found, but not sure how detailed their testing was.

My personal experience is charging to 4.2v doesn't hurt the batteries very quickly. I have some packs that are almost 10 years old that were left at full charge most of their lives and they still have good capacity. Running packs below 30% DOD routinely is what kills them fast.
 
spinningmagnets said:
"...an interesting discovery on dendrites turned up unexpectedly with Nissans experiment on Leaf owners, they gave half the study group a fast charger and half a normal one for a year, then took the batteries away for investigation. The slow-charged batteries had more growth. It seems that high-charge currents grow them one way, high-discharge the other, so...if you charge the same as you discharge, you would have near zero net growth..."

There's something missing here.

If the author purports to compare fast-charge and fast-discharge to slow-charge and slow-discharge (which he implies in the last line) - how do they get the Leaf owners to match their charge rate to discharge rate? You can certainly charge at different rates, but it would seem difficult to force drivers to discharge at similar rates. (Unless you have a load bank or a V2G system or something.)

It would be an interesting result, but based on the talks I went to at AABC last week, fast charging is always worse.
 
ZeroEm said:
As a leaf owner this is good information. Charge at level 2 and worried that I should be slow charging to prolong battery life.
As far as ebike batteries we should be charging at upper range of amps that the battery can handle?
So, if ebike 18650 battery (e.g. 25R NMC A-grade) can handle a continuous discharge of 15amps (MCD rating 20A) then to prevent dendrite buildup you'd recommend charging at a 10amp rate if the average discharge is 10-15amps instead of charging at only a 3-5amp rate to minimize dendite buildup?

Assuming the average 18650 ebike battery discharge is only 10-15 amps is it better to use an 18650 cell with a MCD of 10-15amp rating instead of an 18650 high energy dense cell with a MCD rating of 20-30amp ... to minimize dendrite buildup?
 
I don't have any definitive answers, but this claim is intriguing enough that I felt it deserved its own thread, for future "search terms".

Most ebikers I know charge slowly to reduce heat in the battery. Then as we all know, many ebikers like hard acceleration frequently.

If this proves to be a good hypothesis, it would not be hard to design a charging circuit that occasionally spikes the charge rate for a second or two, but not long enough to cause too much heat.

Only time will tell. I will keep an eye out for any new developments.
 
spinningmagnets said:
If this proves to be a good hypothesis, it would not be hard to design a charging circuit that occasionally spikes the charge rate for a second or two, but not long enough to cause too much heat.
If that is true then with your accumulated knowledge do you presume it better to have a spike in voltage toward the end of the charge cycle or at the very end of the charge cycle (or possibly before recharging an 18650 battery as well as toward the end of a charge cycle)? Every charge cycle or only every 3rd to 5th charge cycle?

Also, if true would it then make more sense that by doing so would be helpful with a high energy dense 18650 battery with the result to "regen" a high energy cell battery (if only temporary) that's more prone to self-discharge than say 2500mAh cells with only a 10amp MCD rating.
 
At this stage, I am just guessing. That being said, I wouldn't perform any amp spikes near the end of charge.

This begs the question, is the formation of dendrites affected by the voltage at all, or will the amps remain the only main suspect?

Thanks for mentioning regen. I hadn't thought about that. Battery packs might still age and lose capacity, but if frequent use of regen prevents the worst of dendrite formation, it could extend battery life, and also reduce the fire-hazard.
 
This one will take study or time to prove out. Don't think leaf or ebike can charge as fast as we pull from the batteries. Do use regen on both. Spiking or surging charger might be a better idea for ebikes. If this study is true then the EV makers will add this feature at some time to extend the life of the traction packs.

EV makers have a setting for 80% charge and start reducing power if the battery gets close to 20%, don't think they would add this if not needed, like brakes. I will stick to the 80-20 rule.
 
Back in the Ni-Cd days, there were "pulse chargers" which pumped out pulses of current that were several times higher than the average. Heating is from the average current but the higher current pulses were somewhat effective at preventing dendrites. Charge rate has other effects, such as the crystal size formed on the plates. Not sure if this is relevant to lithium batteries.
 
The Battery Doctor said:
text
Will your Li-ion batteries last twice as long if you charge them to 4.1V? https://www.youtube.com/watch?v=l0a6kikmS98
I Get Medieval on the science behind 4.1V DOD Research: https://www.youtube.com/watch?v=HhivZn_u5c0
Does Limiting DOD lead to longer life? - Lithium battery myths debunked Part III: https://www.youtube.com/watch?v=eGSJ64Cl_iY

Hi there Battery Doctor. I know your YT channel because I watched some of your videos at past. Nice productions. keep editing for education! Thanks
And the last three videos are very creative!

About 4.1V charging gets twice cycles.. is a amplified scenario, similar to marketing, but in a good way.
There are some good advantages of being under cell limits (assume limit 4.2V)

1st - fully charged battery should not be stored or being all times fully charged because cause more static stress to cells (prove
this)
2nd- its not proved scientifically that 4.1V gives twices cycles, but sure better for cells. thats why most EV's are optimizing
battery usage curve from 20% to 80%, soo should be a good rule..
3rd- Imagine you have a Tesla to drive 100Km everyday, if range is 600Km, it's 1/6 , soo only need to recharge everyday around
20%charge. U could go at 50% SOC battery (40%-60%) everyday keeping at center voltage point. repeat U DON'T need to be fully
charged all time! (U are driving a battery hazard on wheels, so the less percentage charge U have the more safe to roads and to
yourself! Think twice before fully chargin any EV! )

Have a nice time.
 
Battery Doctor is a nice dude doin a great production work plus technical education!
I don't know him, but sure his content is educational @ his channel
We all give him 5 stars! *(invoque Battery Doctor electric God) *edited
 
batteryGOLD said:
*invoque
electric Doctor God is now sleeping, check India time 05:52
I don't know his nationality
Country:
Sweden
https://endless-sphere.com/forums/memberlist.php?mode=viewprofile&u=55042
 
Have not agreed with him much and try to keep that to myself. Was referring to before his latest 12 posts he posted once 18 months before that and once in 2019. Now there is nothing wrong with that. Just saying he maybe gone for a year or so.

Getting back to the thread!
 
Back
Top